dimensional modeling and doing Sentiment analysis using R Script

[yellow_days]

Hello,

I would like to try a simple Sentiment analysis with a R Script.

I created a dimensional model but im not sure how to handle the user comments. I hope someone can give me some advices.

I thought about just to create a simple User Dimension (with name, address, age)

and a fact table with some measures.

I would like the integrate a R-Script (R Script visual) and let the sentiment run on the comments. But right now Im not sure where I should store the comments.

I thought first about store them in the fact table but I also read long text attributes should not be stored in the fact table, because of performance reasons. Now my question: can i store them in a dimension table and still run the R-Script and get the right results?

[Anonymous]

HI @yellow_days,

If your tables contains accurate relationships, it is possible to analysis related dimensional table columns.

Since I'm not familiar with dimensional datasource, I'm not so sure if above method works on all dimensional models.

Notice: current power bi not support analysis many to many relationships.

BTW, you can also post this to analysis service related forum to know how r script work with dimension data model.

Regards,

Xiaoxin Sheng